JOURNAL OF CATSLYSIS 52, 462-471 (1978) Catalytic Hydrogenation of Cyclohexene 3. Gus-Phase and Liquid-Phase Reaction on Supported Palladium E. E. GONZO AND M. BOUDART’ Department of C?Lemical Engineering, Stanford University, Stanford, California 94305 Received September 22, 1977; revised January 16, 1978 The hydrogenation of cyclohexene was studied in both gas and liquid phases on six different palladium catalysts supported on silica gel and alumina between 264 and 308 K at hydrogen pressures from 4 to 89 kPa, cyclohexene partial pressures from 0.46 to 1.5 kPa, and cyclohexene concentrations from 0.15 to 3.0 M. Under these conditions the turnover frequency is inde- pendent of particle size (1.5 to 10 nm) in both liquid and gas phases, of the nature of the support, and of the nature of nonaromatic solvents (methanol, n-heptane, ethyl acetate, and cyclo- hexene). The order of reaction was one-half with respect to hydrogen for the reaction in both liquid and gas phases and zero with respect to cyclohexene partial pressure (gas phase) or with respect to cyclohexene molar concent’ration (liquid phase). At the same temperature and hydrogen pressure, the turnover frequency was about twice as high in the gas phase as in the liquid phase. INTRODUCTION for the hydrogenation of cyclic olefins. In The mechanism of the hydrogenation of particular, with supported metals, it is of alkenes on metal catalysts has been the interest to determine the effect of metal subject of many investigations (1). In 1934, particle size on turnover frequency. We Polanyi and Horiuti (2) proposed a mecha- have reported recently the lack of such an nism for the hydrogenation of ethylene on effect for the hydrogenation of cyclohexene, the surface of metallic nickel. This mecha- in both gas (10) and liquid phases (11) on nism still appears to give a good representa- different supported platinum catalysts. The tion of many aspects of catalytic hydro- present paper reports similar work on genation and related reactions on metals of supported palladium. Among previous in- group VIII, particularly nickel, platinum, vestigations of the hydrogenation of cyclo- palladium, and rhodium. hexene on supported palladium catalysts Information obtained from studies using (12-16) only one (12) gives the surface area isotopic hydrogen both in exchange reaction of palladium metal. At any rate, earlier with saturated hydrocarbons (3-5) and in results will be discussed below together the hydrogenation of alkenes (6-9) has with our own results. contributed particularly to the present state of understanding of these reactions. REACTION IN GAS PHASE On the other hand, kinetic data, especially Expev+imental Apparatus area1 rates, are relatively scarce, especially The gas-phase reaction of cyclohexene 1 To whom queries concerning this paper should with hydrogen was studied in a Pyrex be sent. batch-recirculation system (17) under con- 462 OOZl-9517/78/0523-0462$02.00/O Copyright Q 1978 by Academic Press. Inc. All righta of reproduction in rtny form reserved. CATALYTIC HYDROGENATION OF CYCLOHEXENE 463 ditions designed Qo obtain kinetic data free TABLE 1 from all transport influences. The reactor Koros and Nowak Criterion0 was a glass chamber 25 mm in diameter and about 15 mm high wit,h a medium-grade Catalyst Dispersion Turnover frequency fritted disk to support the catalyst in G) (set-I) amounts between 3.5 and 10 mg. A glass 273 K 283 K coil above the rcact,or served to preheat or precool the reaction gases. The reactor was 3.7574Pd/SiOz 55 0.72 1.36 immersed in a Dewar t’hrough which was 0.57y0 Pd/SiOl 60 0.76 I .42 flowed a 50% v,‘v solution of water and o See Ref. (25). Conditions : hydrogen pressure, ethylene glycol, from a t,hermost,at,ed bath. 10.64 kPa; cyclohexene pressure, 1.0 kPa. The temperature was measured with a thermocouple inserted int,o a well extending was used after passage t,hrough a glass coil through the reactor wall to a point just at 75 K. above the center of the fritted disk. The gases were circulated by a noncont,aminat- ing gas-recirculating pump (18). A flow rate of 960 cm3 min’ was used in a syst’em The amount of palladium in the six sup- of 2.34 liters total volume, and 23-cm3 ported catalysts described below was deter- samples were withdrawn at approximately mined by atomic absorption. It is reported S-min intervals for chromatographic anal- as a weight percentage. A catalyst from ysis. A bypass in t,he system made it Engelhard Industries, 4.557 Pd/A1203, possible to int,roduce and mix the reactant was reduced at 773 K in hydrogen, out- gases in the loop while t)he cat,alyst in the gassed at that temperature, and cooled react,or was still in pretreatment stages. In down to room temperature. Then it was order to facilitate sampling, another bypass exposed to air prior to use. The dispersion of permitted samples to be taken without palladium was determined to be 0.21 by interruption of the flow over t#he catalyst. the titration method of Benson et al. (19) The samples were obtained with a Varian- which was used for all other catalysts as Aerograph X.1-210 six-way sample valve well. Values of dispersion are given in and were analyzed in a Ovarian-Aerograph Table 2. A 2.S4% Pd/SiOz catalyst was A-90-I’ chromatograph equipped with a prepared by impregnation using a mea- 3-m column of 70y0 Chromosorb P and sured amount of PdClz dissolved in a small 30% dibut,ylphtalate which operated at quantity of concentrated HCl. The result- 373 K. ing solution was evaporated to boil off Matheson, Coleman and Bell chromato- excess acid and distilled water was added quality cyclohexene was used after purifica- to obtain the required amount of salt tion through an activated alumina column solution to just reach the point of incipient in order to remove peroxides and traces of wetness of the support (Davison silica gel, oxygen. The A1203 column was breated Grade 02). The catalyst was calcined for 4 h with flowing He at 693 K for 1 h, and a in air at 723 II and reduced in hydrogen at bulb with cyclohexene was then attached 773 Ii. Four catalysts, 0.570/, l’d/SiOz, to the column: after which a freeze-pump- 1.39% Pd/SiOz, 1.45y0 I’d/SiOz, and 3.75% thaw technique was used to bring the l-‘d/SiOs, were prepared by cation ex- cyclohexene t,o t,he storage bulb. The change (20, 21). The first step of prepara- hydrogen, 99.932y0 pure, was further tion was to exchange NHJ+ from a 1: 1 purified by diffusion t#hrough a Milton Roy diluted N&OH solution to the surface of hydrogen diffuser. Helium (99.9959/‘0pure) the silicn gel. The silica gel was 50/100- 464 GONZO AND BOUDART maintain the catalyst saturated with hy- drogen (adsorbed and absorbed) and the rest of the system was out’gassed for 0.5 h. As the catalyst was being reduced, the reactant gases were added and circulated through t’he system, bypassing the reactor. The gases were introduced in the following order: cyclohexene (0.46 to 1.46 kPa), hydrogen (4.52 to 85.3 kPa), and helium (to balance the system at atmospheric total pressure). The circulation of the thermo- static bath solution was started. When the reactor had reached the reaction tempera- 1 I I 1 ture, the reactants were admitted to the 012345 catalyst chamber. The first sample for t/lo3 s chromatographic analysis was taken after 3 min, allowing for thorough mixing within FIG. 1. Hydrogenation of cyclohexene : fractional conversion f vemus time t. Catalyst: 4.88y0 Pd/ the recirculating system loop. The sub- A120a; dispersion : 21% ; Temperature: 273 K; sequent samples were taken at time in- ~HZ/~cyolohexe,,e : 100. tervals ranging from 480 t,o 600 set, depend- ing upon the reaction rate. The react,ion mesh Davison Grade 62. The second step rate was taken as t,he average between the was the cation exchange between the initial slopes of the plot of quantity of Pd(NHs)J2+ ion from the solution and the product (cyclohexane) and quantity of surface NH4 ions. The Pd(NHJ4C12 solu- reactant (cyclohexene) versus Gme. tion was prepared by dissolving PdC12 in concentrated NH?OH solution at 348 K Results and then diluting to obtain a 0.01 M The reaction rate is given as turnover solution. The palladium salt solution was frequency (22) defined as the average num- slowly added to a slurry of the support in ber of molecules of cyclohexane produced 1: 1 dilute NH,OH solution. The slurry was stirred for 2 h at 348 K. The next step of TABLE 2 preparation was to filter the slurry. The filtrate was washed with distilled water and Effect of Palladium Dispersiona dried in vacua in an oven overnight at room Condition Gas phase Liquid phase temperature. The catalysts were calcined Temperature (K) 273 308 in air at 723 K for 4 h and reduced at 773 K Hydrogen pressure 10.84 kPa Atmospheric in hydrogen. Cyclohexene pressure or concentration 1 .O kPa 0.24 M in cyclohoxsne Procedure Loading Support Dispersion Turnover frequency (%) (%) (se0) The standard pretreatment of the cata- GM Liquid lysts listed above consisted of a 3-h out- phi% phase gassing at room temperature (pressure : 2.84 SiOt 11 0.58 6.54 1.3 x 1O-3 Pa), followed by l-h reduction 4.88 A1203 21 0.97 7.95 3.75 SiOz 55 0.72 6.54 in flowing hydrogen (75 cm3 min-‘) at room 0.57 SiOz 60 0.76 6.47 temperature.